摘要
首先对流动性时间序列进行特征分析,根据时间序列自相关和偏相关的特点建立相应的自回归移动平均模型。用LM检验模型的残差是否存在异方差现象;为了很好的描述丛集性的特征,建立了ARMA(1,1)-ARCH(1)模型;然后再对模型的残差做LM和Q统计量的检验,来解释模型的合理性。最后,根据所建的模型求出条件异方差,进而计算流动性风险值。
This paper, at first, analyzes the feature of liquidity time Series. Autoregressive moving average model is built according to the characteristics of correlation and partial correlation. LM is used to test whether the heteroskedasticity phenomenon exists in residual series ;In order to describe the trait cluster better, ARMA （ 1,1 ） -ARCH（ 1 ） model is built too; At the same time, LM and Q statistics are used to test to explain the rationality of Model. Finally, conditional heteroskedasticity is easy to know based on the model, and then the value of liquidity risk can be calculated.
出处
《贵州大学学报（自然科学版）》
2009年第5期4-7,共4页
Journal of Guizhou University:Natural Sciences
基金
江苏省人文社科基金重点项目(06SJB790015)
关键词
流动性风险
异方差
检验
liquidity risk
heteroskedasticity
test
